5 LABOR MARKETS A. LABOR MARKETS, GROWTH AND EMPLOYMENT

Transcription

1 5 LABOR MARKETS A. LABOR MARKETS, GROWTH AND EMPLOYMENT 5.1 A well-functioning labor market and sound labor and employment protection policies can contribute to economic growth by facilitating the reallocation of labor to its most efficient uses, by encouraging investment in human capital, and by making the economy more competitive and better able to adapt to changing circumstances. When combined with enhanced incentives for the jobless to seek new employment, the resulting improved balance between labor supply and demand can also work to increase employment and cut unemployment, and can shorten the average duration of unemployment. These positive effects are more likely to emerge in the presence of macroeconomic stability, competitive product markets and a favorable business environment. Sound employment protection policies can also help workers manage the risks associated with unemployment, lost income, and poor working conditions. 5.2 A flexible labor market becomes especially important in a transition economy such as Serbia, which is facing an impending major restructuring towards different and more productive forms of economic activity. The recent experiences of other transition economies show that: (i) regulations that make hiring and firing difficult or expensive can adversely effect employment by making firms wary of offering a job for fear they will be obliged to provide permanent employment; (ii) high payroll taxes, often earmarked for social programs, raise the cost of labor, thereby undermining competitiveness and creating incentives to evade taxes by hiring workers informally; (iii) government mandates requiring extremely generous bonuses, sick leave, and paid holidays also increase labor costs, adding to the reluctance of employers to create new positions; (iv) rules which prohibit part-time and contract work or compel ablebodied workers to retire at an arbitrary age reduce employment by barring willing workers from the labor market; and (v) bureaucratic red tape can make hiring new workers needlessly slow and expensive Most countries of Central and Eastern Europe faced similar problems while going through the transition from socialism to a market economy and experienced quite similar labor market outcomes. First, the initial decline in employment turned out to be significantly smaller than the contraction of output, particularly in the CIS and the SEE countries. Second, as the private sector was initially slow in creating new jobs, labor shifted to low-productivity services and subsistence agriculture, which, together with labor hoarding by enterprises, acted as shock absorbers. Other common characteristics of labor market performance in the CEEC have included: i) an increase in unemployment to levels much higher than in the EU or OECD; ii) a high incidence of long-term unemployment and continued lengthening of the average jobless spell; iii) high youth unemployment, often combined with low youth participation; iv) high unemployment among those with secondary education or less; and v) the emergence of huge regional disparities in unemployment within each country. 112 See Riboud et al (2002), Haltiwanger et al (2003) and OECD (1994).

2 5.4 Serbia s pre-reform labor market exhibited several characteristics that differ from the typical experience of former socialist economies. These specifics were shared by the other economies of former Yugoslavia, and three such characteristics stand out. First, the labor market was shaped by the particular legacy of the self-management system for enterprises. This led to an even higher level of job protection and overall rigidity, and to huge labor hoarding. Somewhat reformed legislation adopted in 1989 gave employers the formal right to lay off workers. However, this change had a limited impact on actual flexibility, since dismissal procedures remained complex and costly (according to the 1996 Law on Labor Relations, minimum severance payments were set at an extremely high 24 months wages) and firms had obligations which in market economies are met labor market agencies and social funds. During the period of UN sanctions (especially during the first half of the 1990s), firing was prohibited by government decree. Second, while many other socialist countries adhered to a policy of full employment under central planning, open unemployment emerged in former Yugoslavia in the 1960s and increased during the turbulent mid-1980s. Finally, informal employment was for many years tolerated by the government since it was one of the key survival strategies during the 1990s. This led to situation in which about one-third of total employment was in the informal sector. 5.5 Given the importance of the functioning of the labor market for overall economic performance, and set against the background of the performance of other transition economies, the objectives of this chapter are to analyze the recent performance of Serbia s labor market and to identify policy measures that would improve its functioning as a means of promoting economic growth and employment generation. The assessment of labor market policies and performance is a challenging task, especially in an economy undergoing profound structural change, since labor market outcomes can be influenced by a broad range of macroeconomic and structural policies and external factors. In addition, since the benefits of labor market reforms are likely to materialize in the long term, and because Serbia introduced most such measures only very recently (since 2001), their impact may not yet be fully discernible. These factors should be borne in mind when considering Serbia s recent labor market policies and performance. B. LABOR MARKET PERFORMANCE IN SERBIA 5.6 The following section provides an overview of the recent performance of Serbia s labor market, with extensive comparisons with some neighboring countries, as well as with EU and OECD averages. This includes discussions of the levels and structure of participation, employment and unemployment, the size of the informal sector and wage dynamics. Most analyses are based on the Labor Force Survey (LFS) and the Living Standards Measurement Survey (LSMS) keeping in mind the many disadvantages of the data available from the Labor Market Bureau. The LFS is based on international standards and recommendations, which allows cross-country comparison. However, for certain flow data that are not available from the LFS, we used the LSMS. The latter has some minor differences from the LFS, which are elaborated below. 92

3 Participation and Employment 5.7 Serbia s recorded labor force participation rate (the share of employed and unemployed in the total working-age population) has decreased slightly since 2000, but remains relatively high. This indication of the supply of labor is a few percentage points above the average for the CEE countries, but slightly below averages for the EU and OECD. The latter difference is almost entirely driven by levels of participation by youth (age 15-24) which are about 15 percentage points lower than in the EU or OECD (see Table 5.1). When compared with CEE countries, Serbia s participation rate exceeds that of Slovenia, Poland, Hungary, Bulgaria and FYR Macedonia, but falls below that of the Czech Republic, the Slovak Republic and Estonia. Table 5.1 Participation and Employment Rates by Age and Gender, (2002) Gender Participation rate Employment rate Slovak Slovak Age Serbia Bulgaria Hungary Rep. OECD EU Serbia Bulgaria Hungary Rep. OECD All 15 to to to to Male 15 to to to to Female 15 to to to to Sources: LFS, 2002 Serbia; Employment in Europe, 2003 EC; Employment Outlook, 2003 OECD. 5.8 Serbia s recorded employment rate has fallen slightly over the last several years, but remains above the average for the CEEC and close to the level of the better performing transition economies. More specifically, it is above that of Hungary, Poland, FYR Macedonia or the Slovak Republic, but below that of Slovenia, Estonia and the Czech Republic. However, it is below the average levels of the EU-15 and the OECD (where employment is also increasing over time). One reason for relatively high recorded employment appears to be the still very high levels of hidden unemployment or fictitious employment within socially owned enterprises (SOEs). As some respondents perceive themselves as being employed while they are actually redundant within their companies, the official figures may overstate the actual employment rate. This factor was highlighted in the recent Productivity and Investment Climate Survey (PICS), 113 which showed that only 25 percent of companies have an optimal level of staffing. As expected, the situation is worst within the SOEs, of which nearly 80 percent have problems with redundant EU 113 World Bank (2003b). 93

4 labor. This suggests that recorded employment could decline when the intensification of enterprise restructuring converts hidden into open unemployment. 5.9 As in most CEECs, the resumption of economic growth has not resulted in a parallel increase in employment, reflecting labor hoarding. As shown in Figure 5.1, once output for CEE countries started to recover, employment stagnated for several years and only recently began to grow at a slow rate. In the SEE countries, including Serbia, employment continued to decline even after output began to recover in the current decade. CEE countries also tended to experience smaller changes in both output and employment (in particular, a lower decline in output), while Serbia and other SEE countries tended to have a steeper decline of output with a slower recovery afterwards in comparison with the CEE countries. Figure 5. 1 Evolution of Employment and Output, a Figure 5.2 Evolution of Employment Rates, Employment (CEE) Output (CEE) Employment (Serbia) Output (Serbia) Serbia Bulgaria Croatia Hungary Slovak Republic a: (1989=100). Sources: ECA Regional Database WB; Economic Survey of Europe, UNECE; Employment in Europe EC; LFS Serbia Youth (especially young females) have the lowest participation and employment rates. These rates are also much lower than for the same groups in the EU, the OECD and successful transition economies. More specifically, Serbia s youth employment rate is 22 percent, only one-half of the OECD average of 43.7 percent. Serbia s youth participation rate is about 15 percentage points below the OECD average. These relative gaps are even more striking for young females, whose employment rate is less than one-half of the OECD or EU averages. On the other hand, older groups seem to have participation and employment rates above the averages for transition countries, which probably indicates delays in restructuring the real sector One of the most striking and important features of Serbia s labor market is a very high share of employment in the informal sector (Table 5.2). Although the informal sector existed during socialism, it became more prominent in the 1990s and continued to grow in the early years of transition. In fact, calculations based on the LSMS suggest that informal employment accounted for 31 and 35 percent of total employment in 2002 and 2003, respectively. Serbia, where the highest incidence of informal employment is 37 percent in Central Serbia, has one of the highest shares of informal employment in Europe, second only to Bosnia and Herzegovina. 94

5 Table 5.2 Serbia: Formal and Informal Employment by Region, 2002 (percent) Serbia Belgrade Central Serbia Vojvodina Total employment Formal employment Informal employment Source: LSMS Flexible forms of employment (part-time and temporary work) are almost entirely absent in Serbia s formal sector. In 2002, part-time jobs comprised only 1 percent of total wage employment. In contrast, part-time employment represents around 18 and 15 percent of total employment in the EU and OECD, respectively. While low levels of part-time employment are characteristic of many transition economies (the corresponding ratios were only 1.9 percent in the Slovak Republic, 3.6 percent in Hungary and 2.5 percent in Bulgaria), the Serbian levels are particularly low. Similarly, work under temporary or fixed term contracts is fairly insignificant, accounting for only 5 percent of total employment in Part-time jobs have been one of the key factors in increasing employment and participation rates in some OECD countries over the past several decades. These flexible forms of employment can be particularly important in improving the employment prospects for females and youth, the least employed groups in Serbia, and for formalizing existing employment The share of self-employment in total non-agricultural employment in Serbia is one of the lowest among CEE and SEE countries. This ratio is about 5 percent in Serbia, compared to over 10 percent in Slovenia, Hungary, Poland and the Czech Republic, and about 14 percent in the OECD and EU-15. This finding highlights the slow progress in self-employment activity compared to the advanced transition economies of Central Europe, where such activity has played an important role during the transition (Dutz et al., (2001). The relatively low share of employment in small and medium size enterprises, which stood at 29 percent in 2000, also confirms the unsatisfactory performance in this area While the last several years have seen a noticeable shift of employment from the state sector to the private sector, total employment in the private sector remains low. According to the LFS, employment in the private non-agricultural sector rose by 21 percent between 2000 to 2002, while its share in overall employment grew from 24 percent in 2000 to 29 percent in 2002, keeping the state sector as the dominant employer. This trend appears to reflect primarily the privatization of socially-owned enterprises rather than the creation of jobs in the de novo private sector, which probably indicates a still unfavorable business environment (primarily due to barriers to the entry of small businesses), and the relatively slow privatization of large state and public enterprises. 114 This share is still among the lowest in transitional countries, even if one accounts for small firms operating in the informal sector. 95

6 Unemployment 5.15 Contrary to many transition countries where official (registered) unemployment rates usually underestimate the level of joblessness, 115 in Serbia such figures overstate the actual unemployment rate. Various incentives to register lead most of the unemployed to register with the Labor Market Bureau (LMB). The unemployment benefit is not the main reason for registration, as only around 6 percent of the registered unemployed actually receive such a benefit. Moreover, payment of the benefit is in arrears for about six months. The incentives to register come largely from the fact that the registered unemployed can qualify for a broad range of other benefits, such as: health and pension insurance, social assistance benefits, subsidized child care, subsidized local transportation, and access to several active labor policy programs. 116 Box 5.1 provides a more detailed description of the differences between the two main official indicators of unemployment. Box 5.1 Official versus ILO Unemployment Rate The unemployment rate in Serbia in 2002, based on the LFS data, was 11 percent while the official unemployment rate was around 29 percent. The major source of this discrepancy originates from the different sources of data and different definitions of labor force status used in defining the unemployment rate. The LFS uses the standard International Labor Organization (ILO) definition whereby a person is defined as unemployed if the following conditions are jointly met: (i) the person does not have a job; (ii) the person is actively seeking a job; and (iii) the person is available for work. 117 The official unemployment rate is based on the registered number of unemployed with the LMB. The registration data on unemployment do not necessarily meet the above conditions. For example, some individuals may register in order to have access to various types of benefits provided by the LMB (especially health insurance), although they actually have informal sector jobs and, as a result, may not be actively searching for a job. Another reason is that in certain situations registration with the LMB is required to acquire a new job (e.g., graduates who must register first even if they have already found a job). Thus, registration data overstate the actual level of unemployment in Serbia. The second difference between the official rate and the ILO survey-based unemployment rates comes from the discrepancy in the number of employed. More specifically, the official unemployment rate relies on the number of employed persons estimated by the Statistical Office, which does not cover farmers, family workers, on part-time and occasional workers. In addition, the gap between the two measures is a reflection of the large informal sector. Thus, employment based on the LFS appears much larger, as it covers all of those employment categories. The ILO unemployment rate not only provides a more accurate picture of the situation in the labor market, but is also internationally comparable Unemployment in Serbia has steadily increased since the initiation of reforms in late 2000 but remains relatively low compared to many other transition economies. The LFSbased unemployment rate increased slightly from 9.6 percent in 2000 to 11.1 percent in Based on preliminary data from the LFS for 2003, this increase appears to have been sustained in Although Serbia s unemployment rate is fairly average for economies in transition, it is above the levels of well-performing economies such as Hungary, Slovenia and the Czech Republic, and also above the average for OECD and EU countries (see Table 5.4). The surveyed unemployed rate is likely to be somewhat underestimated, as it does not cover redundant workers 115 See EBRD (2000). 116 The latest changes in legislation dealing with employment policies and unemployment insurance (introduced in 2003) reduced the scope and size of unemployment benefits. 117 The third condition is not used in this study, as this information is not available in the LFS. 118 For this analysis we calculated the unemployment rate for Serbia using strict ILO definitions which brings certain differences compared to data published by the Statistical Office. Re-classification of data was done in order to count as employed certain groups which are usually counted by the Statistical Office as either unemployed or inactive. More on this issue can be found in World Bank (2003c). 96

7 who perceive themselves as employed and declare this in surveys. This situation will change when accelerated restructuring converts hidden into open unemployment. Table 5.3 Unemployment Rates, by Age and Gender, 2002 Gender Age Unemployment rate Serbia Bulgaria Hungary Slovakia OECD EU All 15 to to to n.a. n.a. n.a. 6.0 n.a. 55 to n.a. n.a. n.a. 4.9 n.a. Male 15 to to to n.a. n.a. n.a. 5.6 n.a. 55 to n.a. n.a. n.a. 5.3 n.a. Female 15 to to to n.a. n.a. n.a. 6.4 n.a. 55 to n.a. n.a. n.a. 4.2 n.a. Sources: LFS, 2002 Serbia; Employment in Europe, 2003 EC; Employment Outlook, 2003 OECD As in other CEECs, unemployment mainly affects the young. The unemployment rate for those age is almost three times the OECD average for the same age group. While Serbia s total unemployment rate is lower than that in selected transition economies (Bulgaria and the Slovak Republic), its youth unemployment rate is several percentage points higher. This gap is even more striking for young females. Serbia s youth unemployment rate is 3.5 times the overall unemployment rate, compared to about 2.0 in the OECD and EU. Table 5.4 Unemployment Rates in Serbia and Selected Transition Economies, OECD and the EU, Bulgaria Croatia Hungary Serbia Slovak Rep OECD EU Sources: LFS, 2002 Serbia; Employment in Europe, 2003 EC; Employment Outlook, 2003 OECD Unemployment is concentrated among individuals with primary and secondary education (Table 5.5). The unemployment rate for those with a secondary education (13.5 percent), is over twice the level for university degree holders (5.9 percent). This is made even more striking by the fact that persons with a secondary education comprise about two-thirds of 97

8 the labor force. By contrast, those with less than a primary education have the lowest unemployment rate (of 2.9 percent) and their share in total unemployment is low (less than 2 percent). One possible reason is that these persons are more likely to accept job opportunities and are mainly employed in physical work in construction or subsistence agriculture. Table 5.5: Unemployment by Educational Level, 2002 (percent) Unemployment Share in rate unemployment Share in employment All Less than primary Primary completed Secondary completed Higher University Source: LFS The structure of the unemployed by previous labor market status has changed over the last two years, with an increased, but still low, share of job losers compared to new entrants. After remaining stable over the period , the share of job losers in total unemployment increased significantly, from 29 percent in 2000 to 43 percent in 2002 (Table 5.6). However, this fraction is lower than in Croatia (56 percent [Rutkowski, 2003c] and considerably lower than in advanced transition countries (for example, Hungary and Poland) and in OECD countries (around 75 percent), which underscores the rigidities of the labor market and the slow pace of restructuring during the first two years of transition. Table 5.6 The Unemployed by Employment History and Reasons for Unemployment, (in percent) All unemployed New entrants Job losers Those who lost jobs Layoffs Termination of temporary employment Voluntary quits and other reasons Source: LFS The duration of unemployment has been declining, but it remains extremely high, with nearly three-quarters of the unemployed looking for a job for one year or longer. Long-term unemployment is a serious problem, since many of those who wait several years for a job eventually become inactive, which represents a waste of human capital and an additional burden on the social funds. The share of long-term unemployed in the total peaked at 84 percent in 2000 and declined gradually thereafter (Table 5.7). In this respect, Serbia lies between countries such as Bulgaria (64 percent) and Croatia (56 percent) with lower shares, and FYR 98

9 Macedonia with the highest fraction of long-term unemployed (80 percent). 119 The corresponding OECD average of about 30 percent points to the much more rigid labor market for new entrants in Serbia compared to OECD countries. Table 5.7 The Evolution of the Unemployment Rate and Long-term Unemployment, Unemployment rate Share of long term unemployed Source: LFS In 2002, the average uncompleted unemployment spell was 47 months (Table 5.8). Three groups were most affected: women, the middle-aged, and the poorly educated. Women are more likely to be long-term unemployed and have longer unemployment spells relative to men, despite their higher average level of educational attainment. Prime-aged adults require more than twice as long to find a job as do young workers (59.7 months as compared to 24 months for the young). Workers with less than primary education tend to be unemployed for over twice as long as university educated workers, who find jobs, on average, in 33 months. The long average duration of unemployment for most workers may indicate a stagnant unemployment pool in Serbia. 120 Table 5. 8 Serbia: Uncompleted Unemployment Spells by Individual Characteristics, 2002 Mean Median All Gender Male Female Age groups Education Less than primary Primary completed Secondary completed Higher University Source: LFS See Rutkowski (2003a, 2003b, 2003c). 120 Given the fact that unemployment spells represent censored variables due to the availability of LFS stock data, these results should be interpreted with caution. The major disadvantage of the LFS data on unemployment lies in the fact that the LFS data do not represent a panel, so it is not possible to monitor the same individuals over time and investigate changes in their labor market status. This is done using the panel (LSMS ). 99

10 5.22 While the unemployment rate appears to be similar in urban and rural areas, it varies considerably across the regions (Table 5.9). Surprisingly, Vojvodina, which is the most developed and richest part of Serbia, has the highest unemployment rate, at 15.2 percent. The other two regions have much lower rates, especially Belgrade, at 6.4 percent, and Central Serbia, at 10.5 percent. This could be attributed to two factors: (i) the large presence of refugees who are more likely to be unemployed (58 percent of all refugees are located in Vojvodina); and (ii) the faster pace of restructuring (43 percent of all unemployed who were laid off owing to restructuring come from Vojvodina). This may indicate that once restructuring accelerates in the whole country, total unemployment may increase significantly, possibly to levels above those in Vojvodina. Wage Dynamics Table 5.9 Unemployment Rate by Location/Region (percent) Unemployment rate Urban (city) 11.4 Other 10.7 Regions Belgrade 6.4 Central Serbia 10.5 Vojvodina 15.2 Source: LFS The adjustment in the labor market to declining output during the 1990s occurred in the form of lower real wages, with relatively little change in labor force participation and employment. One good representation of this adjustment is from 1999, when real GDP and real wages declined by 18 percent and 21 percent, respectively, while the unemployment rate rose by only 1 percent. At the same time, the labor force participation rate increased from 68.1 percent in 1998 to 69.4 in Such outcomes reflected a continuation of the previous practice under which the state as the dominant employer was reluctant to fire redundant workers. Another form of adjustment to lower levels of economic activity has been non-payment of wages. Some estimates for 2003 showed that over 120,000 employees were not receiving wages, or were receiving wages with long delays, sometimes reaching several years This pattern changed dramatically with the initiation of reforms in late 2000, after which recorded real wage growth rose to well above real GDP and productivity growth (Figure 5.3). 121 By cutting labor demand, this may have contributed to the observed increase in unemployment, (i.e., had wage increases been more moderate, growth could have come with better employment outcomes). Measured in euro terms, average net monthly wages grew very significantly from about 60 euros in early 2001 to about 180 euros in early This reflected not only real wage growth, but also the real appreciation of the dinar over the same period. As 121 The upward trend in real wages also reflects an increase in the coverage (in terms of more employees) and the scope (inclusion of lunch, transportation, etc.) of the wage data. 100

11 shown in Table 5.10, the average euro value of monthly wages is now significantly above the levels in 60 neighboring Romania and Bulgaria. 50 This could undermine the competitiveness of Serbia s 40 economy and erode its ability to 30 attract needed foreign investments. 20 On the other hand, the increase in 10 wages to relatively high levels is a phenomenon observed in most 0 former Yugoslav republics (data for -10 FYR Macedonia and Croatia are -20 presented in Table 5.10), and probably reflects a common heritage of labor market institutions and practices from the former SFRY, as Sources: IMF and Serbian Statistical Bureau well as the high average wage level (over US$500 per month) of the early 1990s. Table 5.10 Net and Gross Monthly Wages in Euros, 2003 FYR Bulgaria Romania Croatia Macedonia Figure 5.3 The Evolution of GDP, Productivity, and Wages, Real GDP growth Productivity growth Growth rate of real net wages Serbia, 2001 Serbia, 2002 Serbia, 2003 Net Wages Gross Wages Source: National statistical offices; calculations into EUR based on official exchange rates as published by central banks. Labor Force Flows 5.25 The entire labor market, including both the formal and informal sectors, shows relatively good flexibility as measured by flows between employment, unemployment and inactivity. 122 This relatively good average performance is primarily driven by a very flexible informal sector, which comprises about one-third of total employment. The informal sector generates the majority of flows it absorbs redundant workers, employs those who were previously inactive, exhibits a high probability of job loss, and allows flexible forms of employment. Only 63 percent of those employed in the informal sector maintained the same status one year later, compared to 82 percent in the formal sector. The informal sector is not only more flexible itself, but annual labor flows between the formal and informal sectors are also significant. Among those workers from the formal sector who remained employed in 2003, around 18 percent were absorbed by the informal sector. By contrast, over a third of the workers in the informal sector were integrated into the formal sector within one year. 122 Labor force flows are analyzed using LSMS data on the same individuals for 2002 and As within-year transitions between different labor force states cannot be captured using the above data, they underestimate actual labor movements during the year. 101

12 5.26 The probability of finding a job in Serbia is relatively high by the standards of transition economies. 123 According to the LSMS data, almost one-half of the unemployed found a job within one year. 124 This proportion is higher than in Russia, Poland and the Czech Republic, all countries with relatively high flows from unemployment to jobs among transitional countries. However, over one-half of the previously unemployed who found a job within the year are in the informal sector. Another important characteristic is that job-to-job transitions are considerable in Serbia, as 9 percent of workers change jobs within a year. This figure is much higher than in Poland and Lithuania, were it amounts to between 4 and 6 percent of employment. Youth (especially those without previous work experience) and those with a lower educational attainment have the lowest probability of finding a job. Flows from inactivity to employment are also high, although the majority of these are movements absorbed by the informal economy. Table 5.11: Serbia: Transition Probabilities from Employment by Age, Education and Employment Type, 2003 (percentage of population in 2002) 2003 Employed in 2002 Employed Unemployed Out of the labor force Total Employed Age Education Less than Primary Primary Secondary Higher/university Type of ownership of employer Non-private Private Type of employment Formal Informal Source: LSMS 2002 and The probability of becoming unemployed is low relative to both transition and OECD countries, but the young and the poorly educated face the highest risk of joblessness. Around 3 percent of Serbian workers who were employed in 2002 became unemployed in This is about the same rate as that for Polish workers during the mid-1990s 123 See OECD (2003). 124 There are certain differences between the data from LFS and LSMS. For example, the share of long-term unemployed differs between flow and stock data. Around one-quarter of the unemployed were long-term unemployed in 2003 using flow LSMS data, while this fraction was over two-thirds based on the LFS of There are two possible explanations for such discrepancies. First, stock data from LFS may overestimate the proportion of long-term unemployed in overall employment. When asked about the duration of their job search, people tend to neglect short spells of employment (e.g., casual, temporary and/or informal jobs). They may also ignore spells out of the labor force, when they were not actively looking for a job, or were not available for a job. The second possible reason is that the 25 percent figure is biased downward owing to the selection bias, the long-term unemployed being more likely to exit the sample (non-response, etc.) than those looking for a job for less than a year. 102

13 and Russian workers in the early 1990s (Bell, 2001), and is also in line with the average for EU countries. 125 Young workers (aged under 30) are more than twice as likely to move to unemployment as middle-aged workers. Workers with only a primary or secondary education face the highest probability of becoming unemployed, a feature also observed in other transition and market economies However, a part of this low risk of moving to unemployment appears to reflect significant flows to inactivity. About 15 percent of those who were employed became inactive one year later. This can reflect the discouraged worker effect. Over one-quarter of the unemployed (26.9 percent in 2002) withdrew from the labor market one year later. It can also reflect the low marketability of skills among those losing their jobs owing to restructuring (e.g., older redundant employees of SOEs going directly into retirement). Table 5.12 Transition Probabilities from Unemployment by Age and Education (percentage of population in 2002) Employed Unemployed Out of the labor force Unemployed Age Education Less than primary Primary Secondary Higher/University Source: LSMS 2002 and Summary of Labor Market Characteristics 5.29 The above analysis shows that Serbia s labor market shares most of the typical characteristics of transition economies. This includes an increase in unemployment in the early years of transition, a high incidence of long-term unemployment, high unemployment among youth and those with a secondary education or less, and the emergence of huge regional disparities in unemployment. Serbia s labor market seems to have a relatively good level of flexibility measured by flows between employment, unemployment and inactivity. However, a disproportionate share of these flows is generated in the informal sector, which accounts for about one-third of total employment and is growing over time The increase in unemployment in Serbia has been noticeably moderate to date compared to that in other transition economies. This phenomenon appears to be associated 125 OECD (2003) discusses transition probabilities for EU as a whole (actually only for the 11 EU member countries for which data were available). This report also provides more detailed data for part-time and full-time employed in the EU. The probability of those with part-time jobs becoming unemployed is 4.7 percent, while the analogous figure for those with full-time jobs is 2.7 percent. 103

14 with the slower pace of enterprise restructuring. Experience from other countries that have already progressed with the transition (and even from Vojvodina) show that once restructuring starts, both the level and structure of unemployment may deteriorate. At the same time, Serbia has to date experienced similar qualitative trends in employment as other ECA countries. The eventual recovery of output came not with a rise but with a further fall in employment This analysis reveals a labor market which is dual in several dimensions, the most striking being between the formal and the informal sectors. The relatively satisfactory overall performance of the labor market reflects the average of a poorly functioning and relatively rigid formal labor market which offers high rates of statutory employment protection, and a very flexible and growing informal labor market which offers no employment protection at all. High levels of informal employment reflect the tolerance of such practice throughout 1990s, tax evasion, inefficient tax collection etc. The chances of both finding and losing a job are much greater in the informal sector. The formal sector (and within it especially SOEs and companies privatized under pre-2000 programs) has yet to respond fully to the more flexible legislation mainly due to unchanged corporate governance and institutional behavior with respect to the labor market. However, this is not a sustainable solution, since informal sector companies have limits on their ability to develop and grow. In addition, those employed in the informal sector have little social security but are financed by contributions from the formal sector Another important duality is between younger (age 15-24) and older workers. The low youth participation rate shows that many young people are not even looking for work. The high average level, but relatively short average duration, of youth unemployment suggests that young people are able to find some jobs (probably low quality or short-term seasonal jobs) but cannot hold onto them for long. This is confirmed by the noted high probability of becoming unemployed. In contrast, older workers have high recorded employment and participation rates, lower levels of unemployment, but very long durations of unemployment once they lose their jobs Women have much lower participation and employment rates than men. While a similar pattern is observed in most economies in transition, the difference is much larger in Serbia (although a few countries such as Bulgaria, Hungary and the Slovak Republic have gaps which are almost as large). Interestingly, the ratio of male to female participation rates (1.3) is still below the OECD average (1.35). Unlike in most other economies in transition, and despite lower levels of labor force participation, the unemployment rate for women in Serbia is much higher than for men. The ratio of female to male unemployment rates is 1.5, a figure only matched by the Czech Republic, although not much higher than the 1.3 average for the EU-15. In contrast, several economies in transition (Hungary, Romania, Estonia and Latvia) have higher unemployment rates for men than for women. In Serbia, women are also more likely to be longterm unemployed and tend to have longer unemployment spells than men The labor market also shows a duality between poorly educated and well-educated workers. People with a higher education have a low unemployment level, have better chances of keeping their job, have higher wages, and spend the least time looking for work. Those with a secondary education are in a much worse situation, are faced with the highest risk of losing a job and have the highest unemployment rate. While it is hard to split the data between those with 104

15 general and vocational secondary education, the highly specialized and outdated vocational secondary education system limits labor mobility Finally, the relatively flexible labor market in the private sector contrasts with a rigid labor market in the budgetary, state and socially-owned sectors. Like the informal sector, the private sector creates most of the flows between unemployment and employment (and vice versa), provides much more flexibility (in terms of offering temporary and part-time jobs), provides wages that are more closely linked to the results of work, and is less burdened with redundancy. On the other hand, the state sector provides higher job security but at the expense of lower productivity and efficiency. Also, the role of trade unions is much higher in the state/socially owned sector. With these differences in mind, the functioning of the overall labor market will change with privatization and public sector restructuring, which will shift a majority of the employed to the private sector In addition, while there are concerns about wage data, Serbia appears to have moved from emphasizing the primacy of sustaining employment through control of real wage growth in the 1990s to allowing more rapid real wage growth at the expense of employment growth since This has led to a less equal distribution of the costs of transition between insiders and outsiders and has contributed to jobless growth. By raising unit labor costs, this trend also works to erode Serbia s competitiveness and may endanger macroeconomic balances. C. KEY CHALLENGES 5.37 The key challenges for policy makers include: (i) moving employment from the informal to the formal sector and reducing the duration of unemployment by improving the flexibility of formal labor markets and by continuing reforms that will expand the share of private sector activities; (ii) sustaining the existing (high by regional standards) overall level of employment and participation; (iii) improving the quality of jobs and overall living standards by facilitating the reallocation of labor to more efficient sectors; and (iv) dealing with the specific problems of youth and the less well educated. Meeting these challenges will require synchronized policies in the areas of labor legislation, the institutional framework for the labor market, severance and unemployment benefits, and active labor market measures. These policies need to be complemented by reforms in macroeconomic policy (including tax policy and administration), privatization and private sector development, restructuring, and education at all levels. The remainder of this chapter begins with a discussion of direct labor market reforms, followed by a brief discussion of reforms outside the labor market. Improving the Functioning of the Formal Labor Market 5.38 Serbia s significant recent progress in reforming rigid labor market legislation to meet the requirements of a market economy needs to be preserved and selectively deepened. This is because more modest employment protection for all is preferable to a system which provides high statutory protection for the few while leaving many others with no protection at all. A new Law on Labor, adopted in December 2001, provided more flexibility in both the hiring and firing of workers (see Box 5.2). These changes are the basis for the efficient 105

16 reallocation of labor so as to increase the productivity of existing and new enterprises. Annex 2 shows indexes constructed under the World Bank project Cost of Doing Business, where a common methodology allows cross country comparisons. 126 This exercise shows that Serbia and Montenegro (data are not separated by republic) has most values close to averages for the transition economies of Europe and Central Asia, but that it maintains terms of hiring and conditions of employment which are more rigid than those in OECD countries. This suggests that the main cause of rigidity in Serbia s labor market is no longer the labor legislation. The recent improvements need to be maintained, and further improvements in flexibility should be considered, especially in the areas identified in the above analysis. Otherwise, foreign and local investors may be reluctant to take part in privatization or to start new business. Also, the government as a large employer might face large fiscal costs with respect to hiring and firing of its employees. The main remaining weaknesses in the functioning of Serbia s formal labor market must be found in other factors that influence institutional flexibility, such as social partnership, minimum wages, unemployment benefits, active labor market programs and tax policy. These factors are explored in the following paragraphs. Box 5.2 Key Innovations Introduced by the Labor Law of 2001 The labor law and other related laws adopted during the previous three years have significantly improved the legal and institutional environment for Serbia s labor market. This is primarily reflected in simplification and reduced costs related to hiring and firing processes as well as the better regulation of working conditions. Hiring has become much easier, faster and less costly since the employer is no longer obliged to announce vacancies, the trial period is shortened, and flexible forms of employment (i.e. part-time and term jobs) are allowed. This should enhance employment prospects, especially for youth and women. Employers can terminate the labor contract if the need for carrying out a certain task ends due to technological, economic or organizational changes. Severance payments were reduced from between 24 and 36 monthly salaries in the previous Law to between two and five monthly wages, depending on job tenure. This will greatly reduce disincentives to hire new workers. Conditions for work are better regulated and terms are harmonized with most transition economies. Limits on overtime work, the minimum duration of annual leave, as well as maternity leave are all precisely defined and in line with regional standards. Minimum wages are set by mutual agreement of the government, unions and employers. Collective agreements have become less restrictive while the representation of unions and employers is more precisely defined. The general and special collective contracts are binding only those employers who are members of the employers association that is the signatory to the collective contract or who have joined the association subsequently. According to the previous legislation the general and special collective contracts were applied directly and were binding upon all the employers in the republic or in a branch. The minister in charge of labor may expand the application of the special collective contracts to include non-members of the employers association. The minister is obliged to request the opinion of the signatories to the collective contract before deciding on such expansion. These changes need time to show their full impact, which depends on reform of the institutional setting of the labor market and on reforms in other sectors. One already observable result is a decline of registered unemployment since mid-2003, the first such decline in over ten years. At this stage of economic development, when the restructuring of large SOEs as well as the attraction of FDIs is so important, flexible labor regulation is much needed. 126 World Bank (2004c). The Employment Laws Index covers three areas: flexibility of hiring, conditions of employment, and flexibility of firing (see Annex 5.2 for cross-country comparisons and for more details on how the index is constructed). Each sub-index takes values between 0 and 100, with higher values implying more regulation. The Employment Laws Index is an average of the three indices. 106

17 5.39 Social partnership has not functioned properly in the initial phase of transition. Social partners are not well institutionalized and some are weakly represented. In addition, the legitimacy of certain social partners is questioned. In general, unions provide workers with a collective voice, increasing their bargaining power relative to management. They can help improve work safety conditions, but can also impose efficiency costs. 127 In addition, if employers are not represented properly, the outcomes of tripartite negotiations might worsen the business environment and lead to wage pressure, slowing the growth of output and employment. 128 In Europe and Central Asia, membership in trade unions was mandatory under socialism. In the early years of transition, trade union members still comprised the majority of the labor force the average union density for the transition countries was 58.1 percent. By the late 1990s, union membership had been further eroded, with only one out of three workers still being a union member. In Serbia, the strong presence of trade unions is evident only in recently privatized companies (66.4 percent) and state and socially owned companies (83 percent of all employed are union members). According to PICS, only 6.2 percent of workers within the new private sector are members of a trade union. In the future, it will be important not to broaden the legally mandated coverage of collective agreements, as this could erode recent important improvements in labor market flexibility 5.40 The further growth of minimum wages relative to average wages could negatively affect employment, especially among young and less skilled workers. Minimum wages are in place in virtually all transition economies, but their generosity has declined over time. While minimum wages can boost the earnings of those low salaried individuals who manage to retain employment, they can also lead to unemployment where the minimum wage is above the marketclearing level and is thus binding. Overly high minimum wages can particularly reduce the chances of being hired for young or low-skilled workers. 129 Thus, minimum wages should be kept at a modest level which will not adversely affect the employment prospects of those who are already the most disadvantaged in Serbia s labor force, namely, youth and those with limited education and low or inappropriate skills. Serbia s current Labor Law determines that minimum wages shall be established by mutual consent of the representatives of the government, trade unions and the association of employers. This principle is in line with the ILO Minimum Wage Fixing Convention, and should be maintained. Recently, the national minimum wage was 38.8 percent of the average net wages in 2002 and 39.4 percent in This is relatively high in comparison with other transition countries. Further increases should be avoided and a slight reduction of this ratio could be considered The recent shift away from legal/administrative employment protection (including through lower minimum severance payments) is positive, but needs to be complemented by well-targeted, effective and affordable programs of financial and active support for job losers. Protecting employment by directly restricting layoffs discourages the creation of new jobs and actually hinders the needed reallocation of labor to higher productivity uses which would lead to growth and sustainable employment in profitable enterprises. In contrast, programs which provide financial and other direct support can protect redundant workers while speeding 127 Rashid and Rutkowski,( 2001). 128 For example, employers should not be represented by the Chamber of Commerce, which is inconsistent with ILO conventions. 129 OECD (1999). In Hungary, the minimum wage was increased in 2001 and 2002 by 60 percent and 25 percent, respectively. A study by Kertesi and Köllö (2002) estimated that this reduced the employment of unskilled workers by 4 percent and 6 percent, respectively. The impact on total employment in 2001 and 2002 was a decrease of 2.5 percent and 4 percent, respectively. 107